Partial Least Squares Discriminant Analysis‐Based Decision Support System for Porous Ceramic Matrix‐Based Burner

ABSTRACT A partial least squares discriminant analysis (PLS‐DA) classification model is first developed for the decision support system in a porous ceramic matrix (PCM) based burner. The PCM‐based burner is numerically solved for the generation of 121 pairs of gas and solid temperature profiles. The data are divided into four different classes, signifying four distinct regimes of operation of the PCM‐based burner, based on the values of the extinction coefficient and convective coupling. The data are then used for the development of the classification model. The developed classification model is then used to correctly classify the operation regime of the PCM‐based burner for 11 new samples. Very high values of sensitivity (1.00), specificity (0.98), and precision (0.96) were obtained for class 3, under the no noise case. The classification model is also explored for two different assignment criteria (Bayes and max) and noisy cases (2% and 5%). Very high values of the classification parameter were obtained for 2% noise case. For 5% noise case, even though the parameters of the classification model were poor, the plots can help identify the corresponding class easily for new samples.